Introduction

Top Abstract Introduction Methods Results Discussion References

Several commonly used general anesthetics, sedatives and anxiolytics (barbiturates, steroid derivatives, alcohols, benzodiazepines and inhalational anesthetics) possess the ability to modulate GABA_A receptors in the central nervous system (Lambert et al., 1995). GABA is the most widely distributed and abundant inhibitory neurotransmitter in the central nervous system. It activates either the chloride-linked GABA_A or the second-messenger-coupled GABA_B receptor (Yeh and Grigorenko, 1995). There is a substantial amount of evidence to suggest that supraspinal GABA_A receptors are important sites for the pharmacological action of many of the nonvolatile and volatile general anesthetics and sedatives (Jones et al., 1995). GABA has also been implicated in the spinal cord control of nociception (Nadeson et al., 1996), but the precise role of the GABA_A receptor in analgesia remains unresolved.

Molecular biological studies have identified the GABA_A receptor as a macromolecular complex. It consists of five subunits with many similarities to the nicotinic acetylcholine receptor complex. Five types of subunit have been described: alpha, beta, gamma, delta and rho, and multiple variants for each of these classes have been identified (alpha, 3beta, 2gamma, 1delta and 2rho). Cloning studies have shown that a number of combinations of these subunits may play a role in GABAergic neurotransmission in the central nervous system (Olsen and Tobin, 1990; Tobin et al., 1991; Wafford et al., 1993). The GABA_A receptor complex contains a number of distinct binding sites for ligands, including GABA, benzodiazepines, barbiturates, picrotoxin and t-butylbicyclophosphorothionate (Sieghart, 1992).

Propofol [2,6-diisopropylphenol (Diprivan)], a sterically hindered phenol derivative, has become used widely as an intravenous anesthetic. This drug is unrelated structurally to any other general anesthetic and is the most short acting of the commercially available intravenous agents. Studies have suggested that propofol produces anesthesia by acting at GABA_A receptors (Jones et al., 1995; Sanna et al., 1995). A study by Hales and Lambert (1991) showed that propofol potentiated the amplitude of membrane currents elicited by the local application of GABA. These results are consistent with biochemical data showing that propofol increases [³H]GABA binding and GABA-induced Cl flux while decreasing [³⁵S]t-butylbicyclophosphorothionate binding in rat brain through a bicuculline-sensitive mechanism (Sanna et al., 1995). The early evidence indicated that modulation of GABA_A receptors by propofol resembled the action of other general anesthetics such as barbiturate and steroid derivatives. However, in recent studies, Concas et al. (1991, 1992) identified a binding site for [³H]propofol in rat brain that differs from either the steroid or barbiturate recognition sites

Studies have shown that positive modulation of the GABA_A receptor in the spinal cord can cause antinociception in vivo (Edwards et al., 1990; Nadeson et al., 1996). No in vivo studies investigating the role of propofol and GABA_A receptors in spinal cord have been reported. In the present study, we investigate the interaction of propofol injected intraperitoneally with GABA_A receptors at the level of the spinal cord. We used two GABA_A antagonists (bicuculline and SR-95531) injected i.t. to investigate this interaction. These results suggested the same spinal GABA_A receptors were responsible for i.p. propofol antinociception as those involved with spinally mediated antinociception after i.t. midazolam (Goodchild et al., 1996). Spinal delta opioid receptors are also involved with midazolam-induced antinociception. We therefore investigated the involvement of spinal cord delta opioid receptors with propofol-induced antinociception by attempting to suppress the effects with the delta-selective opioid antagonist naltrindole.

	Methods

Top Abstract Introduction Methods Results Discussion References

This work was carried out with the permission of the Monash University Standing Committee on Ethics in Animal Experimentation (SCEAE Project No. 93017). In all experiments, attention was paid to ethical guidelines for the investigation of experimental pain in conscious animals (Zimmerman, 1983)

Surgical techniques. Male Wistar rats (weight, 180-200 g) were anesthetized with halothane in oxygen-enriched air (FiO₂ = 0.4) and a Portex catheter (i.d., 0.28 mm; o.d., 0.61 mm) was implanted under aseptic surgical conditions into the lumbar subarachnoid space to lie adjacent to lower lumbar and sacral segments of the spinal cord as previously described(Edwards et al., 1990).

Test for recovery from sedation. Twenty rats (180-200 g; 10 with and 10 without i.t. catheter) were tested for the presence of the righting reflex and competency to perform the Rotarod test. These tests were used to assess the time taken for full recovery from sedation after an i.p. injection of 40 mg/kg propofol (Diprivan; ICI, Melbourne, Victoria, Australia). The rats were naive with no previous exposure to the Rotarod test. They were placed on the Rotarod accelerator treadmill (Ugo Basile 7650 accelerator Rotarod, Ugo Basile, Varese, Italy) set at the minimal speed for two training sessions of 1 to 2 min at intervals of 30 to 60 min. After this conditioning period, the animals were placed onto the Rotarod at a constant speed of 25 rpm. As the animal took grip of the drum, the accelerator mode was selected on the treadmill (i.e., the rotation rate of the drum was increased linearly at 20 rpm every minute thereafter). The time was measured from the start of the acceleration period until the rat fell off the drum; this was the control (pretreatment) performance time for each rat. The maximum running time was 30 sec. This test was performed on each rat four times with an interval of 30 min between each run. The mean performance time was calculated as an average of the last three control performance times.

Nociceptive tests. Experiments were performed on 38 rats with chronically implanted lumbar subarachnoid catheters. The animals were placed in a restrainer that was covered to exclude distracting sights and sounds. Nociceptive thresholds were measured with ECT and the TFL as previously described (Edwards et al., 1990; Nadeson et al., 1996).

Measurement of antinociceptive effect of propofol. ECT and TFL were assessed every 5 min until three stable consecutive control readings had been obtained. An i.p. injection of propofol 40 mg/kg (n = 10) was then administered, and rats allowed to recover from the effect of propofol for 40 min (the mean recovery plus 1 S.D. assessed by the Rotarod test in the previous group of age- and weight-matched rats). Nociceptive values were then measured every 5 min for 60 min (fig. 1, protocol 1).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Two protocols for the experiments in this study. Protocol 1 was used to study the antinociceptive action of 40 mg/kg propofol alone. Protocol 2 was used to assess the interaction of the same dose of propofol with GABAA receptors at the level of the spinal cord. X1-X3, control period before drug; ECT and TFL measurements. Y1-Y3, propofol (40 mg/kg) alone; ECT and TFL measurements. Z1-Z3, nociceptive thresholds measurements after i.t. injections of antagonists bicuculline, SR-95531 and naltrindole

Experiments using antagonist. The involvement of spinal cord GABA_A and delta opioid receptors with the antinociceptive effect of i.p. propofol was investigated in 38 rats with chronically implanted i.t. catheters. Experiments were performed on each rat once daily, up to 6 successive days. These animals were placed in a restrainer as above, and protocol 2 in figure 1 was followed. Nociceptive thresholds using ECT (neck and tail) were measured every 5 min until three consecutive stable readings had been obtained at each skin site. Propofol (40 mg/kg i.p.) was then given, and the rats remained in the restrainer. After full recovery from propofol-induced sedation, the residual antinociceptive effects were assessed before and after i.t. administration of bicuculline methiodide (Sigma Chemical, St. Louis, MO), SR-95531 (Research Biochemicals, Natick, MA) or naltrindole HCl (Research Biochemicals) (see protocol 2). The volume of the antagonist injected i.t. was 5 µl. This was chosen so that spread of the drug in the cerebrospinal fluid was minimized and its effect would therefore be restricted to the lumbosacral spinal cord; all antagonists were also dissolved in a slightly hyperbaric 6% dextrose solution for the same reason. To confirm this restriction of the drug effect, ECT was measured at both skin sites every 5 min. A change in the tail ECT after i.t. drug without a change in the neck ECT would indicate that the action of the drug was confined to the caudal segments of the spinal cord responsible for tail innervation.

12

11

12

11

11

8

Controls. Experiments conducted in which 5 × 10¹¹ mol bicuculline (n = 6), 5 × 10¹¹ mol SR-95531 (n = 6) or 5 × 10⁸ mol naltrindole (n = 6) was administered i.t. but without prior i.p. propofol. This was done to determine whether these compounds had any effects on ECT or TFL when given alone.

Statistical analysis. All statistical comparisons between groups were made using Student's unpaired t test or one-way analysis of variance as appropriate. A value of P < .05 was considered statistically significant. ED₅₀ values were calculated by applying a nonlinear regression to each line using the STATISTICA computer program.

	Results

Top Abstract Introduction Methods Results Discussion References

All animals exhibited positive lidocaine tests after each experiment, indicating that the drugs injected down the catheters had been given into the i.t. space. No rat showed overt signs of neurological damage during the series of experiments by either loss of motor power or spontaneous occurrence of anesthesia in the tail.

Recovery from sedation after i.p. propofol. Propofol (40 mg/kg i.p.) caused no loss of righting reflex, but it did cause decreased performance in the Rotarod test. Table 1 shows the time (mean ± S.E.M.) taken to lose and recover performance in the Rotarod test after i.p. propofol (40 mg/kg) for a group of normal rats (n = 10) and rats with i.t. catheters (n = 10). There was no significant difference in the recovery times between rats with and without i.t. catheters. In pooled data for the two groups, full recovery from sedation measured by the Rotarod test occurred at 36.3 ± 1.7 min (mean ± S.E.M.) after i.p. propofol. In subsequent experiments, all nociceptive readings were taken 40 min (36.3 min + 3.6 min, which is 1 S.D. from the mean) after i.p. propofol.

Antinociceptive effect of propofol. Figure 2 shows time-response curves for the antinociceptive effect of propofol assessed using ECT (measured at neck and tail) and TFL. Propofol produced a rise in electrical current threshold in the tail (R = 2.26 ± 0.27 × control; mean ± S.E.M., n = 6) and in the neck (R = 2.09 ± 0.27 × control; mean ± S.E.M., n = 6) without affecting TFL (%MPE = 9.77 ± 9.81; mean ± S.E.M., n = 6). The antinociception produced by propofol assessed using ECT was sustained for 60 min (the end of the testing period allowed under ethics committee license) after full recovery from propofol-induced sedation.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Time-response curves for the antinociceptive effects of 40 mg/kg propofol i.p. Experiments were performed as in protocol 1. Arrow, time of i.p. injection of propofol. Black bar, period of sedation; , neck (ECT); , tail (ECT); , TFL (%MPE). Values shown are mean ± S.E.M. (n = 6).

Reversal of propofol antinociception by antagonists. All three antagonists produced suppression of the residual antinociceptive effects of propofol in the tail but not the neck. Figure 3 shows one such response as a time-response curve, the mean ± S.E.M. from six experiments in which bicuculline (5 × 10¹¹ mol) suppressed the antinociceptive effect of propofol assessed using ECT at the tail skin site but not the neck.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Time-response curves for the antinociceptive effects of 40 mg/kg propofol i.p. and the effect of 50 pmol bicuculline i.t. Experiments were performed as in protocol 2. Arrows, time of injection of 40 mg/kg propofol i.p. and 50 pmol bicuculline i.t., respectively. Black bar, period of sedation; , neck (ECT); , tail (ECT); , TFL(%MPE). Values shown are mean ± S.E.M. (n = 6).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Antagonist dose-response curves. The suppression by bicuculline (, ), SR-95531 (, ) and naltrindole (, ) of the ECT antinociceptive effects of propofol (40 mg/kg i.p.; closed symbols) and i.t. midazolam (47 nmol; open symbols). A, ECT tail skin site. B, ECT neck skin site. ED50 values calculated for each curve (STATISTCA) are bicuculline vs. midazolam, 4.3 × 1011 mol; bicuculline vs. propofol, 3.4 × 1011 mol; SR-95531 vs. midazolam, 2.2 × 1012 mol; SR-95531 vs. propofol, 2.4 × 1012 mol; naltrindole vs. midazolam, 2.9 × 1010 mol; naltrindole vs. propofol, 4.7 × 1010 mol.

Controls. Intrathecal injections of bicuculline (5 × 10¹¹ mol), SR-95531 (5 × 10¹¹ mol) or naltrindole (5 × 10⁸ mol) alone or the vehicles (10% intralipid for i.p. propofol and 6% dextrose for i.t. injections) caused no change in antinociceptive response assessed by either ECT or TFL values at the doses used (table 2).

View this table: [in this window] [in a new window]   TABLE 2 ECT response (r) and TFL (%MPE) in a groups of rats given %6 dextrose i.t., intralipid 10% i.p. or i.t. bicuculline (5 × 1011 mol) i.t. SR-95531 (5 × 1011 mol) or i.t. naltrindole (5 × 108 mol)

	Discussion

Top Abstract Introduction Methods Results Discussion References

Propofol is a commonly used anesthetic that has been shown to enhance GABA-mediated synaptic inhibition in a number of neuronal systems (Hales and Lambert, 1991). In our experiments, an i.p. injection of propofol (40 mg/kg) caused rats to become sedated but not lose consciousness. Full recovery from the sedative effects as measured by the Rotarod test took ~35 min, after which the animals had resumed all normal grooming and feeding behavior. Nociceptive tests were carried out 40 min after the i.p. injection of propofol to assess whether any residual (subanesthetic/sedative) concentrations of propofol caused any antinociceptive effects. Although the rats in which nociceptive thresholds were measured could not be subjected to the Rotarod test while in the restrainer, it is likely that they recovered from the sedative effects of i.p. propofol (40 mg/kg) similarly to the rats in the previous group; they were obtained from the same batch and were the same age and weight. In addition, it was possible to observe them while they were in the restrainer, and they showed no signs of sedation. The residual brain concentrations of propofol (40 mg/kg) that were probably present were therefore not sufficient to cause a noticeable change in conscious level.

After full recovery from sedation, the residual concentrations of propofol produced a rise in both neck and tail ECT values with no change in TFL. This antinociceptive effect was long lasting, and no appreciable reduction was noticed for 50 min after time had been allowed for full recovery from sedation. Because this antinociceptive effect occurred in the absence of any detectable anesthesia or sedation, it can be suggested that the systemic concentration of propofol needed to maintain antinociception is far lower than that required to induce or sustain anesthesia or sedation. However, full recovery and return of nociceptive thresholds to base-line values had occurred by the following day before the next experiment in the series.

In the present study, it is clear that propofol acts similarly to i.t. midazolam in that it has no modulatory role in the nociceptive pathway activated by noxious heat stimulation. Previous studies have shown that the results from a number of different nociceptive tests can yield contrasting results. For example, midazolam, a benzodiazepine, also causes spinally mediated antinociception when given i.t. in our model (Nadeson et al., 1996); this drug causes a rise in tail ECT without a change in TFL. This is in contrast to i.t. fentanyl, which causes antinociception when assessed by both tests (Serrao et al., 1989a). It seems likely that the ECT activates different nociceptive afferents to those stimulated by the TFL test.

An i.t. injection of the GABA_A antagonist bicuculline or SR-95531 caused a dose-dependent reversal of the antinociceptive response to ECT in the tail but not in the neck. This differential effect suggests that there was no rostral spread of the antagonist in the CSF to the cervical levels of the spinal cord or higher because no effect was seen at the neck skin site. It can therefore be concluded that the antinociception produced by propofol is mediated through GABA_A receptors at the level of the spinal cord because both GABA_A antagonists could reverse the tail ECT rise. The antagonists did not reverse the antinociception after i.p. propofol at the neck skin sites. Because the antinociception in the tail site could be reversed by the spinal application of the antagonist, we may conclude that the effects were mediated by spinal cord mechanisms.

In experiments using the delta opioid antagonist naltrindole, a similar dose-dependent reversal of the antinociceptive effect of propofol measured by ECT was observed at the tail skin site but not at the neck site. This result suggests that propofol antinociception is also mediated by a spinal delta opioid receptor.

When using a particular antagonist, the dose-response curves for suppressing the antinociceptive effect of midazolam given i.t. and propofol given i.p. are the same. This was true for all three antagonists used. When two different antinociceptive agents are suppressed by the same antagonist with the same dose-response relationship and the same ED₅₀, this implies that the drugs are acting either directly or indirectly at the same receptor (Mackay, 1994); that is, the two agents both bind to the receptor or cause release of an endogenous neurotransmitter that binds to the same receptor as the antagonist drug.

This present study suggests that the antinociceptive effect of i.p. propofol is mediated by the same GABA_A receptors as previously described for i.t. midazolam even though the drug is given by a nonspinal parenteral (Goodchild et al., 1996). Propofol, like midazolam, does not bind with delta opioid receptors. However, both drugs must cause a release of endogenous opioid peptides that bind with spinal cord delta opioid receptors that were blocked by naltrindole. It is unclear whether propofol binds directly with the spinal cord GABA_A receptors or exerts an indirect action. This could occur by activating brain systems with pathways descending to the spinal cord that then interact with spinal cord GABA_A receptors. To elucidate these questions further, experiments using i.t. administered propofol are required. At present, this is not possible because no suitable vehicle for the drug is available for i.t. injection.

Antinociceptive effects of propofol have been reported in humans (Anker-Moller et al., 1991), and there has been some suggestion from in vitro studies that this action occurs via modulation of GABA_A receptors. Our studies show that propofol causes antinociception in rats through involvement of the same GABA_A and delta opioid receptor observed for midazolam. These mechanisms suggest a possibility for potentiation of conventional mu opioid analgesia. Midazolam, however, has to be given i.t. to produce this effect (Yanez et al., 1990). The current evidence suggests that this potentiation occurring at the level of the spinal cord may be achieved with propofol without the necessity of invading the i.t. space. This may be useful in humans, in whom one is always wary of the possibility of neurological damage.